09.9.19.23:59: THE JANA LANGUAGE: ASPIRATES FROM K-LUSTERS

In my last post, I showed how Jana developed aspirated consonants from consonant-liquid clusters. I left out a technical detail: r would become ʁ before becoming h.

Here's a very precise account of the development of aspirates. p stands for any voiceless stop or affricate and b for any voiced stop or affricate:

Aspirates also came from two other types of clusters:

1. k-clusters

Note that kk remains kk and doesn't become an aspirate kh.

I was inspired by the aspirates of Korean (ph, th, chh, kh) which probably partly came from k-clusters.

2. Aspirate clusters

Korean aspirates also came from h-clusters: e.g.,

'big' : 크- khɯ- < *hkɯ- < *hɯkɯ-

(The h-k- is attested in 鷄林類事 Kyerim yusa [1103].)

Hence I propose clusters with fricatives (f, s, sh, x, h but not the approximant-like v*) as a third source of Jana aspirates. H stands for any fricative in the chart below:

I don't know what ske in Skear (a Kasgen mountain name) means, but its Jana cognate, if any, might be khe:

ske > hke > khe

The changes above only apply to stops and affricates.

Fricative-fricative clusters become geminate fricatives:

Fricative-sonorant and sonorant-fricative clusters become voiceless sonorants or sh:

k-sonorant clusters either become voiceless sonorants or kh:

In short, if you see an aspirate or a voiceless sonorant in Jana (i.e., any consonant written with h other than ch and sh), you are seeing a simplication of an earlier cluster.

I suspect that Tangut and Old Chinese aspirates and voiceless sonorants may have similar origins.

*Jana (and Kasgen?) v may be like the Dutch labiodental approximant w which is IPA [ʋ].

09.9.18.22:59: THE JANA LANGUAGE: WHERE DID THE ASPIRATES COME FROM?

The Jana language is the only language on Hadanus with aspirated consonants: e.g., ph, bh. In my previous post, I gave a preview of their origin:

Træotnu > Thæonnu

Earlier Jana once had syllable-initial clusters like Dwar. However, current Jana only allows three kinds of initials:

simple: e.g., p, b

aspirate: e.g., ph, bh

geminate: e.g., pp, bb

Like the geminates, the aspirates also came from earlier clusters:

The shift of Cr to Ch in Jana was inspired by Muong, a relative of Vietnamese. Muong kh is partly from earlier *kr.

9.19.12:15: The r to h shift in Jana is like the English substitution of h for Norman r: e.g.,

Hick(s), cognate to Richard

Hodge(s), cognate to Roger

9.19.12:19: hC represents a preaspirated consonant. Such consonants are less common than regular Ch aspirates. I don't know of any language that contrasts preaspirated and aspirated consonants, so I merged the two into a single aspirated category in stage 4.

09.9.17.23:59: THE JANA LANGUAGE: WHERE DID THE GEMINATES COME FROM?

Although I glanced at the Wikipedia article on geminates (long consonants) when writing last night's post on Jana, I didn't notice this passage that David Boxenhorn pointed out to me:

There are very few languages that have initial consonant length; among them are Pattani Malay, Chuukese, a few Romance languages such as Sicilian and Neapolitan, and many of the High Alemannic German dialects (such as Thurgovian). Some African languages, such as Setswana and Luganda, also have initial consonant length—in fact initial consonant length is very common in Luganda and is used to indicate certain grammatical features.

Although I've briefly looked at Setswana, I have no idea where its geminates came from. I don't have time to research the other languages tonight, but I know that Proto-Austronesian (the ancestor of Pattani Malay and Chuukese) and Proto-Indo-European (the ancestor of Sicilian, Neapolitan, and Thurgovian) are not reconstructed with geminates. My guess is that geminates in those Austronesian and Indo-European languages originated from earlier nongeminates: e.g.,

unaccented syllables losing their vowels: C₁VC₂ > C₁C₂ > C₂C₂

clusters: C₁C2 > C₂C₂

Austronesian languages typically don't have initial clusters, so the latter explanation may only work for the IE languages.

(9.18.00:59: I totally forgot about the initial geminates of Nha Heun [Ferlus 1971 in Sagart 1999] which come from earlier clusters:

mm- < tm- or km-

nn- < pn- or kn-)

I've never heard initial geminates, but I suspect they sound like Korean reinforced consonants which are romanized as if they were geminates (e.g., pp, ss, tt, kk).

Korean reinforced consonants generally come from earlier clusters: e.g.,

ppyŏ 'bone' < spyə

ssal 'rice' < psʌl

ttae 'time' < pstay

kkul 'honey' < pskul

(9.18:00:43: Added examples mostly from Lee and Ramsey [2000: 285].)

Medial geminates can also come from earlier clusters: e.g.,

Japanese Nippon 'Japan' < earlier *Nitpon

Pali kamma 'karma' < Sanskrit karma

(9.18.0:03: Many more Pali examples here.)

The geminates above are the results of assimilation: one consonant (C₁) becoming like another (C₂).

The geminates of Jana could also result from assimilation: e.g.,

Salvesh > Savvesh

Træotnu > Thæonnu

I'll discuss the shift of tr > th [formerly tt] next time.

09.9.16.23:59: THE JANA LANGUAGE: ASPIRATES, VOICELESS SONORANTS, AND GEMINATES

Robinson wants me to make the Jana language of Hadanus more distinct. Ideally a reader of his fiction should be able to look at a word and guess whether it's Jana, Kasgen, Dwar, or Urdreh. What could make Jana unique? I could just decide to add clicks to it so it can sound like, say, Zulu, but I'd rather draw upon what Robinson already established starting in 1990.

In my first Jana post, I extrapolated a whole set of aspirated stops and affricates on the basis of th, a consonant unique to Jana.

This morning I thought of going even further with new consonants written with h. What if Jana had voiceless sonorants (nasals and approximants) like the Hm- of Hmong or the Lh- of Lhasa? Here's a table of the expanded Jana consonant inventory which is the largest of all known languages on Hadanus. New consonants are in bold.

(h)v is, strictly speaking, a fricative, but can be voiceless or voiced like an approximant, whereas all fricatives can only be voiceless.

Jana may or may not have the consonants in parentheses. (h)ny is like Spanish ñ. x is like Spanish j rather than [ks]. hny, ny, hng, and x would be unique to Jana.

Jana is the only language on Hadanus with geminate (long or 'doubled') final consonants: Thall. Jana also has medial geminates: e.g., Enissa, Samanna. What if most or even all Jana consonants could be geminated?

Note that some spellings of geminates are abbreviated: e.g., hnny instead of hnyhny.

One could make Jana really exotic by allowing long initial consonants. I don't know of any language with them, though some may sound like Korean reinforced initials: e.g., ttal 'daughter'.

Existing Jana consonant clusters could be replaced by geminates: e.g.,

Swithgar > Hvitthar

Traeotnu > Ttaeonnu

Aspirates, voiceless sonorants, and geminates should be used sparingly. There is no language in which

ph hm bb

are more common than

p m b

However, they could be used to 'spice up' a Jana name that might be mistaken for a name in another language.

Lastly, both Jana and Dwar have more vowels than vowel letters, necessitating the use of vowel letter combinations to represent single vowels: e.g., in Jana

eo = [ə]

ae = [ɛ]

ao = [ɔ]

These could be replaced with Scandivanian letters:

Note that only the Scandinavian letter shapes are being recycled, not their phonetic values.

Examples:

nixheon > nixxön or nixxøn (or nixxeon)

Maerhu > Märru or Mærru (or Maerru)

Slaw > Hlå (or Hlao)

The vowel letter combinations of Dwar could mostly be replaced by vowel-accent combinations. Compare the table I wrote in June with this table:

Examples:

Koen > Kőn

Haen > Hèn

The letter ő is taken from Hungarian in which it represents long [øø] rather than short [ø]. I chose ő to resemble a mix of ó and ö.

Perhaps Jana could have Danish/Norwegian-style ø æ å while Dwar has acute, grave, umlaut, and double acute accents: e.g., ó, ò, ö, ő. My only concern is typeability.

09.9.15.23:57: THE JANA LANGUAGE: VOWEL INVENTORY

The Jana language of Hadanus has at least five vowels - the usual ones:

However, there are hints that it could have more:

ui in oluin

ee in Tanjeel

ei in Mei. Selimei, Amanei, Galei

eu in Aireund

eo in nixheon

ae in Maerhu

aeo in Traeotnu

2. Written vowel-glide sequences that might represent single vowels:

ay: Maytran

aw: Slaw, Malaw

3. The letter y in a vowel-like position in Myria which probably doesn't have a consonant cluster [myr] before [i]. (I have never seen a language with such clusters.)

I could assign a single vowel to each written vowel sequence but this would result in a vowel system almost as complex as that of Dwar, and I don't want the Jana vowel system to resemble that of any other language on Hadanus. I'd like Jana to be partway between the extremes of Urdreh with only four vowels and Dwar with twelve basic vowels (not including nasal and retroflex variants or diphthongs). Hence I propose a system of nine basic vowels like Thai:

y is halfway between i and u. If a syllable contains y without any other vowel letter (a e i o u), then y is a vowel rather than a consonant. [yɨ] could be spelled yy. The use of y for this vowel is taken from the romanization of the similar Russian vowel ы.

eo, as its spelling indicates, is a neutral vowel between e and o and sounds like Urdreh e (which is [ə] and not [e]).

ae, as its spelling indicates, sounds like both a and e.

The use of eo and ae to write [ə] and [ɛ] is based on the Revised Romanization of Korean.

ao, as its spelling indicates, sounds like both a and o. Although this was originally spelled aw, I've changed it to parallel ae and to use a digraph I already used for Dwar (though its sound value is slightly different).

Vowel sequences ending in -i and -u are treated as diphthongs: e.g.,

Mei [mei] (one syllable; Me'i [me i] would be two syllables)

Aireund [ai reund] (two syllables; A'ire'und [a i re und] would be four syllables)

Apostrophes are used to indicate syllable breaks whenever spellings are ambiguous: e.g., e'o is [e o] but eo is [ə].

Jana could have 16 possible diphthongs:

I regard ay and ee as variant spellings of ei and i.

Any written vowel sequence absent from the above tables consists of a combination of vowels and/or diphthongs: e.g., aeo is not a single vowel or diphthong, but a vowel-vowel sequence: Traeotnu [trɛ ot nu] (three syllables).

Next: Jana consonant clusters.

The Jana language is the fourth major language of Hadanus and the only one I haven't described yet. The Arktikans do not speak and almost nothing is known about the Sabakan language*.

Here's a list of all consonants in known Jana names and words**. Consonants in bold occur in other languages of Hadanus.

This initial list is fuller than lists of attested consonants in other languages of Hadanus because most names and words in existing texts are Jana.

The palatal column (j sh y without ch) is reminiscent of Janhunen's reconstruction of Para-Mongolic which also lacked a ch. Did pre-Jana ch become sh, or merge with sh?

Generally if a language has a voiced affricate, it also has a voiceless affricate: e.g., if it has j, it usually has ch. Thus I think it's likely that Jana also had a ch.

(Classical Arabic is an exception to the j-predicting-ch rule. It lacks ch but developed a j from earlier g. Iraqi Arabic has developed a ch from earlier k.)

Jana is the only Hadanus language with th (presumably pronounced like the English th of through since all known fricatives on Hadanus other than v are voiceless). Since the name Jana is from Sanskrit jana 'person'**, I had been considering reinterpreting th as an aspirated stop and adding two further rows so that Jana would resemble an Indic language and would be different from any of the other languages of Hadanus:

However, none of these consonants with the exception of th appears in the existing Jana lexicon. Therefore I am reluctant to propose such a major addition to the Jana sound system.

I suspect x (only attested in nixheon) is really a cluster [ks] rather than IPA [x] (the -ch of Bach). Thus nixheon would be three syllables [niks he on] and could be respelled as niks'heon (with an apostrophe to distinguish it from a hypothetical niksheon [nik she on])..

I could fill up the v-row with other voiced fricatives like z and zh but they are not attested. Moreover, v and w are in complementary distribution: one appears only where the other does not:

v: syllable-initial: Tevel, Salvesh

w: in consonant clusters: Swithgar

w: syllable-final: Slaw, Malaw

It's also possible that -w actually represents vowel quality in final position: aw = [ɔ] rather than [aw]. I'll expand on this in a post on Jana vowels. If w is a variant of a phoneme /v/ (and a vowel quality indicator?), then the inventory of Jana consonants might be

Next: The Jana vowel inventory.

APPENDIX: Examples of Jana names and words:

Globe-Hurler Prologue: Aireund

GH chapter 1: Belgos, oluin, Arone, Traeotnu, Garon, Swithgar, nixheon, Maerhu, Farlius, Epilegion

GH chapter 4: Enissa

The Fist chapter 2: Noman Thall, Jana, Mei, Dresh, Tevel Slaw

TF chapter 3: Fistun, Ashul Maytran, Armun, Myria, Kain, Heral

TF chapter 4: Tanjeel Elan, Arlen, farli, Kelhan, olum, Samanna, Selimei, Krain

TF chapter 5: Espen, Hamun, Salvesh

TF chapter 7: Malaw, Amanei, Galei, Shun

Jana profile: Elan

Jana role-playing game stats: Thallsend

*The only two Sabakan words revealed thus far are Yushi and Hidei from Globe-Hurler chapter 6.

**I suggested the name Jana to Robinson in 1993. I did not intend the name to imply that the Jana language was Indic or an Indic-influenced language.

After my initial posts on the Kasgen language, Robinson posted a list of Kasgen calendrical terms which did not fit my earlier version of the Kasgen phonemic inventory. This new list incorporates consonant clusters in those terms.

The 16-17 consonants of Kasgen

Glottal stop ʔ- is normally unwritten in syllable-initial position but can be written as an apostrophe to separate two syllables: e.g., a'i [ʔaʔi].

I chose v- instead of w- because of the names Vida and Vidan. Although v is a fricative, it behaves like an approximant w, and it may be pronounced as [w] in nonstandard Kasgen.

The 13 s-clusters of Kasgen

I deliberately included clusters like sb- and sng- to differentiate Kasgen from English. I excluded sf- ss- sh- because I assume that Kasgen has a constraint against initial fricative-fricative sequences, but it might have sf- as a counterpart of sv-.

sb- sd- sg- may be pronounced [zb zd zg]. s- might be [z] before all voiced consonants in some dialects: e.g, sm- as [zm]. [z] does not appear by itself in Kasgen.

sy- may be pronounced [ʃ] before i to clearly distinguish between si and syi. Some speakers may pronounce sy- as [ʃ] before all vowels.

The 41 finals of Kasgen excluding zero for open syllables

I added -st because of the day names Kast and Lest and added -sp and -sk because they are implied by -st.

I added -lm because of the month name Kulm and added other -lC clusters since it is unlikely that -lm is the only -lC cluster in the language.

I don't allow any final clusters with glottals (ʔ, h).

If there is no final -ʔ, perhaps there should also be no final -h. If the Kasgen have no -h, then they would pronounce Urdreh as two syllables [urd re] without -h. (dr- is not a permissible syllable-initial cluster in Kasgen.)

I don't allow any approximant-approximant clusters.

9.13.2:15: If I don't allow -lh or -rh, maybe I shouldn't allow any -l/r-fricative clusters (i.e., -lf, -rf, -ls, -rs). Eliminating them would reduce the number of finals to 38 (37 + zero for open syllables) and the number of possible syllables to 5700 (30 initials * 5 vowels * 38 finals). The 30 : 38 initial : final ratio looked suspicious to me at first until I realized that English has more finals than initials: e.g., -ngths in strengths cannot be an initial.